Research map / Detection and segmentation / Open-set and few-shot detection
Hierarchical Object Representation for Open-Ended Object Category Learning and Recognition
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use independent topics per category instead of topics shared among all categories (ablation) | proposed topic layer · Object dataset 10-fold CV | accuracy: 0.79 -> 0.87 | Helped |
| use topic layer instead of BoW layer (ablation) | proposed representation · Object dataset 10-fold CV | accuracy: 0.79 -> 0.87 | Helped |
| use topic layer instead of pure shape feature layer (ablation) | proposed representation · Object dataset 10-fold CV | accuracy: 0.12 -> 0.87 | Helped |
| voxel size 0.04 instead of 0.03 | topic layer · Object dataset 10-fold CV | avg accuracy (%): 85 -> 81 | Hurt |
| image width 8 instead of 4 | spin images · Object dataset 10-fold CV | avg accuracy (%): 83 -> 83 | No clear effect |
| support length 0.04 to 0.05/0.06 | spin images · Object dataset 10-fold CV | avg accuracy (%): 82 -> 83 | No clear effect |
| dictionary size 50 to 90 visual words | BoW dictionary · Object dataset 10-fold CV | avg accuracy (%): 82 -> 84 | Helped |
| number of topics 30 to 50 | topic layer · Object dataset 10-fold CV | avg accuracy (%): 84 -> 82 | Hurt |
| vary order in which categories are introduced across five random sequences | open-ended system · RGB-D Object Dataset open-ended | GCA (%): 65 -> 69 | Mixed |
| more categories introduced over time | open-ended system · RGB-D Object Dataset open-ended | accuracy | Mixed |
Closest papers by meaning
- Object-Centric Representation Learning for Enhanced 3D Semantic Scene Graph Prediction (NeurIPS 2025 · similarity 0.86)
- Semi-Open 3D Object Retrieval via Hierarchical Equilibrium on Hypergraph (NeurIPS 2024 · similarity 0.85)
- Boosting Segment Anything Model Towards Open-Vocabulary Learning (AAAI 2025 · similarity 0.85)
- Interactive Visual Task Learning for Robots (AAAI 2024 · similarity 0.85)
- Predicate Hierarchies Improve Few-Shot State Classification (ICLR 2025 · similarity 0.84)
- SUGAR: Pre-training 3D Visual Representations for Robotics (CVPR 2024 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.